WORLD CONFERENCE

What's new since last conference? A quick review of the most prominent features included since the last conference held at Budapest.

With a TB-size database, export, restore and migration require downtime measured in days. Businesses working 24×7 — and using the ERP as the platform for their operations — are hit hard by that downtime. Here we see how to achieve migration with almost 0 downtime and just one re-login from the user, at their own comfort level, while the system slowly moves from old to new.

This presentation shares the speaker’s practical perspective on how iDempiere distributions can be upgraded in a sustainable, low-risk way. Drawing on more than ten years of experience with iDempiere and the maintenance of JPiere, it covers how to handle monthly minor upgrades, major-version transitions, plugin compatibility, and upgrade strategies for integrators delivering solutions to end-user companies.

In this session, Hawk Chen, a senior consultant from the ZK Framework team, explores the long-standing and evolving technical synergy between ZK and iDempiere. Hawk Chen will share key milestones from the past year’s collaboration, highlighting critical updates to the framework’s foundation—such as Jakarta migration support and modern theme enhancements—that ensure iDempiere remains robust and future-proof. The presentation will introduce a new suite of open-source components designed to enrich the ERP user experience and demonstrate specialized integration projects that bridge the gap between iDempiere’s core and high-performance enterprise tools. Furthermore, attendees will get an exclusive look at how ZK is embracing the AI era. Hawk Chen will demonstrate cutting-edge AI-assisted development tools, including the ZK DOC MCP and zul Writer, which are designed to significantly accelerate the creation of robust UI extensions and complex business forms. This session is essential for developers and integrators who want to elevate their application's capabilities, streamline their development workflow, and participate in a vision for a next-generation, AI-driven ERP ecosystem.

As database size grows, performance degrades and sometimes reports take longer hours and window loads slow down too. This session shares experience of performance tuning and the results of the same.

Cloud hosting is convenient and expensive. For many organizations running iDempiere, a single bare-metal server is faster, cheaper, and more private. What has been missing is a reliable and scalable way to run one.

Incus is the hosting platform. Incus runs almost everywhere (laptop, desktop, server in a closet, and a server in a rack). One physical machine becomes many isolated system containers, so production, test, and training instances of iDempiere run side by side with independent lifecycles, snapshot and clone in seconds, and a hard boundary between them.

NixOS is the operating system. The entire machine — OpenJDK, PostgreSQL, the systemd unit for iDempiere, the nginx reverse proxy and its TLS termination — is declared in version-controlled files. A rebuild is reproducible, an upgrade is a commit, and a bad change rolls back at the boot menu. There is no undocumented server state, because the configuration is the documentation.

NetBird is the accessibility layer. A WireGuard mesh gives named, authenticated access to each container from anywhere, without exposing iDempiere or PostgreSQL to the public internet and without maintaining a traditional VPN concentrator.

In many business environments, tracking technical changes is not enough. Organizations need to understand how data looked at a specific point in time, preserve the integrity of historical documents, and guarantee consistency across audits, reporting, and retroactive business operations.

The History Records plugin enhances and extends the standard auditing capabilities of iDempiere beyond the traditional AD_ChangeLog mechanism, introducing true temporal data management for ERP master data. It enables iDempiere to manage historical validity of master data transparently and efficiently, allowing users to retrieve not only the current value of a record, but also the exact value that was valid when a transaction or document was created.

iDempiere v14 introduces Extension Management to replace manual OSGi bundle deployment with a centralized registry and safe plugin lifecycle control. This session shows how to configure extension repositories, install plugins with dependency validation, and enable/disable extensions without database risks. Through live demos, you'll see real-world examples of managing REST API, HR, and custom extensions. Ideal for implementors and developers who want predictable, auditable plugin management in production environments.

AI is changing how we write, review, test, and deploy software much faster than most ecosystems can adapt. In this session we look at iDempiere as a “software ecology”: a socio‑technical system made of people, culture, tools, plugins, repositories, infrastructure and community practices. Building on ideas from Adam Bender’s “Software engineering at the tipping point,” we explore what an AI‑first future could mean specifically for iDempiere development.

We will start with a short, practical introduction to systems thinking and “software ecology” and then map those concepts onto the iDempiere world: community repositories, plugins, CI/CD, test coverage, branches and forks, documentation, support channels and governance. Using concrete examples, we will discuss how AI assistance could amplify both strengths and weaknesses: more code, more tests, more tokens, more services and higher operational complexity.

The session will focus on questions that are highly relevant for the project: What would a “10x” increase in change volume look like for iDempiere? Where would things break first? Which cultural and technical practices (testing, code review, standards, plugin quality, documentation) matter most to keep the ecosystem healthy as AI adoption grows?

Attendees will leave with a shared vocabulary to talk about the iDempiere ecosystem, concrete risk areas to watch, and a set of ideas for how our community can prepare for an AI‑first future.

AI tools are rapidly changing how we design, develop, test, and implement iDempiere, however each team experiments in isolation, and we rarely compare approaches. This interactive roundtable brings together developers and technical implementers to share concrete, technical ways they are using AI in the iDempiere ecosystem: from code generation and refactoring, to test coverage, migration scripts, data quality checks, documentation, and deployment automation. The session will start with a short framing of typical AI use cases and open questions (governance, code quality, privacy, licensing), followed by a facilitated “open panel” where participants briefly share how they use AI in their day‑to‑day work with iDempiere. We will collect patterns, tools, and workflows, but also failures and anti‑patterns, and group them into emerging “recipes” the community can build on. The format is intentionally sales‑free, vendor‑neutral, and focused on learning from real experience: what works, what does not, and where we need better community practices and tooling. Participants will leave with a clearer picture of how peers are using AI today, ideas to improve their own workflows, and suggestions for future community initiatives (plugins, guidelines, or shared experiments).

This session looks at what it takes to run iDempiere in production beyond a single server, based on a real, working deployment rather than a theoretical design. Instead of one instance, the setup is built as a resilient, clustered environment: two iDempiere nodes coordinated through a Hazelcast cluster, fronted by a pair of HAProxy load balancers in high availability, and backed by a PostgreSQL database in replication. Plugin delivery is fully automated through a Jenkins pipeline that installs bundles with Apache Felix FileInstall, keeping releases fast and repeatable across environments.

I will walk through the architecture end to end, explaining the reasoning behind each choice, the trade-offs involved, and the real problems that surfaced along the way — from cluster discovery and cache invalidation to session behavior, deployment ordering, and configuration that must survive every release. The goal is to give attendees a concrete, reproducible reference for a business-continuity-oriented iDempiere, with the pitfalls to avoid clearly called out.

The talk also opens an honest discussion on what is not yet solved. Chief among these is transparent, seamless failover for active users: today a node can fail over, but not without disrupting people mid-work. I will summarize what the ZK Enterprise session-failover features and the related iDempiere plugin appear to offer, what remains unclear, and where the community's experience could help close the gap.

This session is for implementers, system administrators, and developers who run, or plan to run, iDempiere in a clustered, highly available way. Attendees will leave with a clear picture of a real production setup, the choices behind it, and a shared set of open questions for pushing high availability in iDempiere further.

iDempiere's data model is its greatest asset and its hardest interface. Every new frontend, portal, mobile app or AI agent ends up rebuilding the same joins, the same tenant filters and the same document-status logic — and usually pays for it in round trips.

This session shows how CloudEmpiere exposed iDempiere through a GraphQL layer that clients can query in a single request, and what it took to make that layer fast rather than merely elegant. We cover the schema design decisions — what to model as graph nodes versus what to keep as explicit operations, and why not every Application Dictionary table deserves to be exposed. We look at the performance work that GraphQL demands and that most tutorials skip: batching to kill the N+1 problem against a normalized ERP schema, projection so the database only reads the columns the query asked for, caching, and query cost limits so a single client cannot bring down the ERP.

Security gets its own section. Tenant isolation is enforced on AD_Client_ID at the application layer, role and window access rules still apply, and authentication runs through Cognito — the API cannot become a side door around iDempiere's own access model.

We finish with the honest comparison: where GraphQL clearly beats the core REST API, where REST remains the better answer, and what the operational cost of running both actually is.

Contributing to iDempiere core is one of the highest‑leverage ways to improve the platform — but success depends on clear scoping, solid design, and a smooth path to merge. This session distills real‑world patterns from recent LTS releases and community workflows into a practical playbook for contributors who want their changes accepted faster and with fewer reworks.

One of iDempiere's greatest strengths is the ability to do most things in many ways. One of the biggest challenges is to know what way is best for you.

An AI agent arrives knowing SQL, Java, and shell — and knowing something about the iDempiere Application Dictionary. Sometimes it guesses correctly. Other times it invents columns, forgets conventions, and goes in circles. The gap is not always model capability. Sometimes the model needs to know: “what is our way of doing things”.

Closing that gap is the catalyst for everything else in this talk. We will give you an example deployment that creates your tenant, organizations, business partners, and artifacts to be able to conduct business in iDempiere with no manual configuration.

While iDempiere's native WebUI offers robust request and project management, its interface isn't always the most intuitive. Enter PRIMHUB: an open-source, multi-platform Flutter application powered by the iDempiere REST API. PRIMHUB transforms these core capabilities into a seamless Help Desk and project management experience, designed to be highly accessible for both internal teams and external customers. This showcase demonstrates a ready-to-use frontend, proving how iDempiere's REST API empowers the creation of user-centric services backed by a powerful core engine.

From transparency to action: review how we spent our resources, celebrate wins, acknowledge gaps, and dream up the next generation of community priorities. A closing session that closes loops while opening new possibilities.

It is Tuesday afternoon and a customer sends a WhatsApp message asking for a quote. By the time someone checks the inbox, the opportunity may already be gone. This is where most customer relationships begin today, yet the operational truth of the business still lives somewhere else: inside the ERP.

This session shows a practical way to close that gap using AI agents and iDempiere. Johan will walk through one real customer message, from the moment it arrives to the moment it becomes a structured, actionable record inside the ERP. The agent reads the message, understands what the customer needs, finds the right Business Partner or creates a new one, and captures the details that matter: what is being requested, by when, and under what conditions. It then prepares a CRM or workflow record that someone can review and act on immediately. By the time someone gets back to the customer, most of the manual work is already done.

This is not about letting AI make business decisions or bypassing ERP processes. It is about removing the repetitive work that slows teams down while keeping iDempiere as the system of record. AI acts as an orchestration layer around the ERP, not a replacement for it. It reads, classifies, and structures information, while people remain in control of approvals, exceptions, and decisions that require business judgment.

Attendees will see the complete flow live, from the incoming customer message to the resulting ERP record. More importantly, they will understand how it actually works and leave with a practical blueprint they can adapt to their own organization, regardless of the AI tools or technology stack they use.

A real case of intercompany operations across Italy, Poland and Spain, with currency, VAT and electronic invoice integration. Creation of the intercompany document starting from a requisition made by MRP.

Every morning, before the drivers arrive, a beverage wholesaler's dispatcher has to load around nine trucks with up to three hundred delivery notes - and be done before the trucks roll. Each delivery note looks like a single line, but hides a dense bundle of constraints: weight, kegs and cases, whether it needs one driver or two, a trailer, returnables, a key kept at the depot, a delivery time window, cash-only customers, which weekdays are even allowed, and which side of the truck a pallet must stand on for a one-way street. It is one of the most demanding screens in the whole ERP - and it is a core point of user experience and usage performance.

This talk shows how we rebuilt iDempiere's tour planning as a fast, information-dense board that thinks like a dispatcher. We'll walk through a working prototype: an at-a-glance card that encodes a dozen attributes without text; instant drag-and-drop dispatching with an optimistic UI; a top-down "load plan" where you pack the truck's cargo bed pallet by pallet and the screen warns you when a pallet sits on the wrong side; and the tour drawn on a free OpenStreetMap with cached routing.

Then we lift the lid: underneath sits a reusable iDempiere building block - a generic "Pinboard" view type and a status layer - so the next hard board is a configuration, not a new plugin. You'll leave with concrete ideas for designing power-user screens, and a pattern for turning a one-off into something the whole community can use.

iDempiere is a good demonstrator for many principles of ERP-systems and general digitalization in business. At DHBW diverse university courses used iDempiere in recent years. The courses ranged from creating and supporting (new) businesses and, In the first years of usage of iDempiere in education, the emphasis has been on systematic understanding of master data and the relations between different document actions to get business smoothly running. The scope of the projects evolved by integrating other technical challenges like layout of printed documents, configuration of workflows, usage of complex products. In other scenarios students worked on projects integrating iDempiere workflows e.g. with IoT-devices, a web-shop, or a robot.

At the moment the goal is to configure an adaptive cost accounting and allocation. After a short elaboration on underlying concepts it is intended to share expertise with other participants of the conference.

We have a process where we have 3rd party business partner like Supplier chaos: AP teams spend 40%+ of their time chasing invoice status, resolving mismatches by phone and email. Customer blind spots: customers have no visibility into order or production status, driving high CSR call volume. Manual bank recon: bank statement matching is largely manual by 2–5 accountant days per month on exception clearing. PLC data is stranded: shop-floor sensor and OEE data never reaches the ERP, production orders close on estimates, not actuals. No AI in the loop: every exception — AP, AR, banking, quality — lands in a human inbox rather than being triaged automatically.

Pharmaceutical is regulated industry which require strict compliant software. iDempiere is successfully implemented in Pharma and health care industries by Logilite

What if every analytics tool you reach for just works on your iDempiere database - cleanly, safely, without re-inventing the same joins? This talk shows how, starting from one idea: shape the data once, in a set of well-crafted views.

The raw iDempiere schema is a wall - hundreds of tables, cryptic codes, mixed units, credit memos with the wrong sign, and a super-user with no row-level safety. So we built a curated BI view layer: pure SQL on top of a live database (or a replication server). It exposes only what matters, resolves codes to words, normalizes currency and units, pre-computes year-over-year columns, and filters every row by tenant using iDempiere's own role rights. It is one single contract every tool can build on. The materialized part can even run on a separate reporting server.

The heart of the talk are two power applications on top of that layer:

• Apache Superset: live dashboards with BI charts by drag-and-drop, with row-level security enforced by the database itself.
• AI Agent: an agent plugin inside iDempiere that answers plain-language questions by writing read-only SQL against the view layer and returns charts, reports or analytic documents - and that structurally cannot leak or change data. It builds reusable patterns for trustworthy self-service analytics for end-users on iDempiere.

Once a second dashboard exists, the same metric has two definitions. Once an AI agent starts answering questions about revenue, it needs to be reading the same number the CFO does. This session shows how CloudEmpiere put a Cube.dev semantic layer over iDempiere: measures and dimensions defined once, in version control, and served over an API to dashboards, applications and agents alike.

We cover why a semantic layer was chosen over querying views directly — tenant isolation enforced in the model rather than in each consumer, and pre-aggregations that turn slow analytical queries over an ERP schema into fast ones. We go through modelling iDempiere's document and accounting structures as cubes, the fan-out traps that silently double your numbers, and how AD_Client_ID becomes a security boundary inside the model itself.

Then we show the consumer side: a custom analytics UI built in Angular and Nx against that API, rather than an off-the-shelf BI tool.